Journal Logo


Malignant Bone Tumors: Limb Sparing Versus Amputation

DiCaprio, Matthew R. MD; Friedlaender, Gary E. MD

Author Information
Journal of the American Academy of Orthopaedic Surgeons: January 2003 - Volume 11 - Issue 1 - p 25-37
  • Free


Primary malignant bone tumors are rare lesions, with fewer than 3,000 new cases per year in the United States. Before the 1970s, management routinely consisted of transbone amputations or disarticulations, with dismal survival rates (10% to 20%). With the development of more effective chemotherapeutic agents and treatment protocols in the 1970s and 1980s, survival rates improved, which allowed the focus of management to shift to limb preservation.1 Computed tomography and magnetic resonance imaging (MRI) allow the precise visualization of the anatomic location of a tumor and its relation to surrounding structures. Preoperative planning has been advanced through the use of these modalities, fostering better patient selection for specific treatment strategies and lowering the morbidity rates of biopsy and subsequent resection.2,3 Currently, 80% to 85% of patients with primary malignant bone tumors involving the extremities (eg, osteosarcoma, Ewing's sarcoma, and chondrosarcoma) can be treated safely with wide resection and limb preservation. Multimodality therapy has increased long-term survival rates of patients with chemotherapy-sensitive tumors to 60% to 70%.1,4 There are a number of options for skeletal reconstruction after bone tumor resection, and it is important to compare the clinical and especially functional outcomes based on type of reconstruction, location of the tumor, and limb-sparing versus ablative surgery.

Limb-Sparing Surgery Principles and Guidelines

There are four basic principles or goals of limb-sparing procedures: (1) Local recurrence should be no greater and survival no worse than with amputation. (2) The procedure, or treatment of its complications, should not delay adjuvant therapy. (3) Reconstruction should be enduring and not associated with a large number of local complications requiring secondary procedures and frequent hospitalizations. (4) Function of the limb should approach that obtained by amputation, although body image, patient preference, and lifestyle may influence the decision.5

Before consideration of limb preservation, the patient needs to be appropriately staged and assessed through a multidisciplinary approach. Some elements of the disease may warrant concern, including relative contraindications to such procedures (Table 1). Multidrug neoadjuvant chemotherapy, popularized first for patients with osteosarcoma by Rosen in the late 1970s, is usually initiated as appropriate after histologic diagnosis and staging. Chemotherapy helps control systemic disease by attacking micrometastases, dramatically increasing overall survival rates.1 Neoadjuvant therapy also “sterilizes” the reactive zone around the tumor by destroying microscopic disease at the periphery of the primary lesion, thus facilitating resection. Additionally, in some patients with a relative contraindication to limb salvage, such as a pathologic fracture in the upper extremity, the use of chemotherapy with a favorable response may allow limb salvage to be considered. However, not all malignant bone tumors (and especially chondrosarcomas) have a viable and effective chemotherapy regimen.

Table 1
Table 1:
Relative Contraindications to Limb-Sparing Procedures

When appropriate, after 8 to 12 weeks of preoperative neoadjuvant chemotherapy, wide tumor resection is performed to establish local tumor control. Achieving tumorfree resection margins is of paramount importance and remains the primary goal in surgical oncology. Most resections are performed through an extensile longitudinal incision, permitting access to the major neurovascular bundle, with complete removal of all biopsy tracts. By definition, a wide resection will include a cuff of normal tissue surrounding the resected specimen. Skeletal defects are large, averaging 15 to 20 cm, reflecting the size of these tumors and the need for negative margins. After reconstruction, muscle transfers may be necessary to provide adequate motor function. Finally, adequate, healthy soft-tissue coverage is essential to prevent early wound complications and subsequent infection. Coverage may require complex local flaps or even free tissue grafts. When indicated on the basis of histologic diagnosis, adjuvant chemotherapy consisting of multiple agents for synergistic activity is continued for 6 to 12 months after wide tumor resection.

After negative tumor margins are obtained, there is often a large skeletal defect requiring reconstruction. Several options are available.6 Patient age, tumor location, and extent of disease narrow the list of appropriate surgical alternatives.

Modular Endoprostheses

Indications and Advantages

Currently available metallic prosthetic systems offer a lightweight, strong, inert means for skeletal reconstruction. Modularity of prosthetic design allows intraoperative flexibility based on the final amount of tissue resected. A rigorous rehabilitation program can be initiated immediately after implantation (usually done with bone cement), allowing early joint range of motion and weight bearing. Prosthetic reconstruction carries a lower risk of deep infection than do allografts, and nonunion is not a concern because there are no osteosynthesis sites. Endoprosthetic use also avoids the risk of disease transmission and immune responses that exists with allograft reconstruction. Longevity, complications, and functional outcome vary by anatomic site, type of prosthesis, and fixation technique.

Complications and Clinical Results

Early complications associated with the extensive nature of most musculoskeletal oncology procedures include wound necrosis/ dehiscence, infection, thromboembolic disease, neurapraxia, and joint instability. Meticulous surgical technique and attention to soft-tissue handling and reconstruction can significantly decrease the frequency of these complications. Late complications include aseptic loosening, infection, joint or prosthetic instability, fatigue fracture of the prosthesis, and wear or dissociation of modular components. Late infection remains the most serious problem because most prosthesis-related complications can be successfully treated with revision surgery. Concerns include the suboptimal attachment of soft tissues to the metallic components.

In an attempt to determine prosthesis and extremity survivorship, Horowitz et al7 reviewed their experience with 93 prosthetic reconstructions over 8 years: 16 proximal femur, 61 distal femur, and 16 proximal tibia. Minimum follow-up was 24 months (mean, 80 months). Prosthesis survival at 5 years was 88%, 59%, and 54% for proximal femur, distal femur, and proximal tibia reconstructions, respectively. The overall event-free prosthesis survival was 63% at 5 years and 36% at 10 years. Aseptic loosening was cause for failure in approximately 20% at 5 years and 30% at 10 years. Limb survival for the entire group was 87% at 5 years and 81% at 10 years. Patients with lesions of the proximal tibia had the longest survival rate, with 93% alive at 10 years. (Histologies included 11 osteosarcomas, 4 malignant fibrous histiocytomas, and 1 chondrosarcoma.) The group as a whole had a survival rate of 72% at 10 years and was composed of a variety of histologies, including 65 osteosarcomas, 13 chondrosarcomas, 10 malignant fibrous histiocytomas, 3 Ewing's sarcomas, and 2 liposarcomas.

Aseptic loosening is the primary long-term concern with this method of reconstruction for tumors around the knee. Whereas allografts successfully stabilize after 3 to 5 years, prostheses begin to exhibit their inherent biomechanical limitations after 10 years. For the current rotating-hinge knee design, reported follow-up is limited to approximately 10 years. Malawer and Chou8 in 1995 showed an 83% survival of prostheses at 5 years and 67% at 10 years. Of 52 patients who survived and were available at 3-year follow-up, only 10 were available at the 10-year followup. They had a revision rate of 15%, infection rate of 13%, amputation rate of 11%, and local recurrence rate of 6%. Overall, 44% of patients had at least one complication. These survivorship data are limited because of the small number of patients but may represent an improvement from the simple-hinge, custom-made prostheses, for which 5-year survival of 80% drops to 53% at 20 years.7,9-11

Of 1,001 patients treated with cemented, custom-made endoprostheses, aseptic loosening was the principal mode of failure among the 210 requiring revision.10 Seventyfour revisions (35.2% of those revised) were done for aseptic loosening. At 10 years, the rates of aseptic loosening were 6.2%, 32.6%, and 42% for the proximal femur, distal femur, and proximal tibia, respectively. The poorest prognosis for prosthesis survival without aseptic loosening was in young patients (<20 years) with distal femoral prosthetic reconstruction in whom a high percentage of femur had been replaced. Experience to date indicates that acetabular loosening rates have been extremely high, simple-hinge prostheses have a higher loosening rate than do rotating-hinge designs, and cemented fixation provides the lowest rate of loosening.9 Most loose prostheses can be revised to improve functioning.

Infection rates range from 0% to 13%.8,9,11 Proximal tibial reconstructions carry the highest risk of infection, as do other regions where softtissue coverage is tenuous. Infection is the most serious complication associated with limb-sparing procedures and is the most common reason for amputation after attempted reconstruction. Rates have decreased, however, with the more common use of rotational or free flaps now readily available through microsurgical techniques.

Joint instability is a major concern in reconstructions about the hip and shoulder. Dislocation rates for the hip range from 10% to 15%.7-9,12,13 Reconstruction of the abductor mechanism and the use of bipolar components have improved stability and function.11

Fatigue fracture of intramedullary stems has become extremely uncommon with the increase in stem diameters, improvements in design, and current metallurgy used during fabrication.11 Dissociation also is rare with modern prostheses; most join with Morse tapers.

Regional Considerations

Prosthesis survival for proximal femoral replacements is generally reported as 77% to 100% at 10 years, falling to 57% at 20 years.9,11-13 Poor abductor muscle function remains a common reason for decreased functional grades. Zehr et al12 reviewed their experience with 33 patients after proximal femoral resection and reconstruction with an allograftprosthetic composite (16 patients) or a megaprosthesis (17 patients). The primary mode of failure for composites was infection, and for megaprostheses, instability. The 10-year prosthetic survival rate was 76% for the composite group and 58% for the megaprosthetic group. Instability occurred in 0% and 28% of the composite and megaprosthetic cohorts, respectively; infection rates were 17% and 6%, respectively. Both groups functioned well, with 87% and 80% functional scores for the respective cohorts. Allograft-prosthetic composites have been shown to have a survival advantage over megaprostheses used for proximal femoral reconstructions.12,14

Overall survival of a simplehinge distal femoral knee replacement prosthesis at 5, 10, and 20 years is 80%, 65%, and 53%, respectively.9,10 Functional evaluation reveals 69% to 93% good to excellent results with less than 10 years of follow-up.7-11,15 Prosthetic survival analysis shows that a higher percentage of femoral bone resected distally is related to a higher risk of prosthetic failure.15 The extent of soft-tissue resection is another important factor. The most common cause of failure for distal femoral prostheses is aseptic loosening. As length of follow-up increases, the rate of prosthetic survival diminishes. Better long-term results are anticipated since simple-hinge designs have given way to modular rotatinghinge systems (Fig. 1).

Figure 1
Figure 1:
A 16-year-old boy presented with knee pain of 7 months' duration and a history of bilateral retinoblastoma in infancy. A, Anteroposterior radiograph of the knee shows an aggressive, eccentric, osteoblastic distal femoral metaphyseal lesion with extensive periosteal reaction (Codman's triangle, arrow). Coronal (B) and sagittal (C) T2-weighted MRI scans demonstrate the heterogeneous lesion, periosteal reaction (white arrow), and soft-tissue extension (black arrow). Open biopsy confirmed the diagnosis of high-grade osteosarcoma. Treatment included preoperative chemotherapy, wide tumor resection, skeletal reconstruction, and postoperative chemotherapy. D, Anteroposterior radiograph after distal femoral reconstruction with a modular rotating-hinge knee prosthesis.

Proximal tibial prosthetic replacement survivorship has been poor because of tenuous soft-tissue coverage and unreliable extensor mechanism reconstruction. Survival rates vary from 45% to 74% at 5 years and 45% to 50% at 10 years.9 Malawer and Chou8 found proximal tibial replacements to have the highest complication and revision rate and worst Musculoskeletal Tumor Society (MSTS) functional scores (Table 2) for any region reconstructed. Wound problems and subsequent infections have decreased with the routine use of flaps but remain frequent problems for reconstructions in this anatomic site. Grimer et al17 reported an initial infection rate of 36% that was reduced to 12% by the use of a medial gastrocnemius flap. Local recurrence was observed in 12.6% of patients and was associated with poor response to chemotherapy and close margins of excision. They also found that 70% of patients at 10 years required further surgical procedures and reported a 25% risk of amputation at 10 years.17

Table 2
Table 2:
Musculoskeletal Tumor Society Functional Evaluation 16

Many designs and techniques are available for proximal humeral replacements. The technique of Malawer and Chou,8 including implantation of a large-segment prosthesis stabilized by static reconstruction with Dacron tape and dynamic reconstruction by muscle transfers, has yielded the best reported results of any site of prosthetic replacement, with an average MSTS functional score of 86.7% (26/30) and no cases of instability. However, there is a wide range of results reported for this site, with instability the primary reason for poor functional outcome.7,9,11 O'Connor et al18 reported on 11 patients who underwent proximal humeral prosthetic reconstruction. Two demonstrated evidence of stress shielding, six had signs of instability, one had a deep infection, and two had ceramic prosthesis loosening or fracture. Four of the 11 patients went on to secondary arthrodesis. Those not converted to arthrodesis were satisfied regarding pain, emotional acceptance, and manual dexterity and were dissatisfied regarding function, positioning of the hand, and lifting ability.

Osteoarticular or Bulk Allografts

Indications and Advantages

Frozen allografts have been used longer than any other tumor reconstruction option. Allografts are favored by some for their potential for longevity because they function as a biologic reconstruction. Incorporation of the allograft by the host is a slow and incomplete process. Osteoarticular allografts permit the uninvolved portion of the joint to be preserved; this approach allows the strongest means of soft-tissue or periarticular ligament reconstruction (Fig. 2). Although associated with more early complications than are endoprostheses, allograft reconstructions stabilize after 3 to 5 years and therefore do better in long-term follow-up studies. Mankin et al19 found that, after 3 years, approximately 75% of grafts are retained by patients and remain successful for more than 20 years. Seventy percent to 80% of patients obtain a good or excellent functional result after allograft reconstruction, although this varies with type of graft, anatomic site, and stage of disease.19 The unpredictable early outcomes with allografts and the frequent need for multiple procedures to obtain a successful end result have led many orthopaedic oncologists to favor modern endoprostheses. With their potential for long-term stability, however, allografts play a key role in younger patients (<20 years), in whom an enduring reconstruction can limit the additional revisions seen with long-term follow-up of patients with endoprosthetic constructs.

Figure 2
Figure 2:
A 17-year-old boy presented with knee pain of 2 months' duration. Radiographs, MRI, and biopsy confirmed osteosarcoma. A, Initial postoperative anteroposterior radiograph demonstrates distal femoral reconstruction with an osteoarticular allograft stabilized with two orthogonal dynamic compression plates. The patient underwent a bone-graft procedure 1 year later to treat an allograft-tohost bone nonunion. B, Anteroposterior radiograph 2 years after reconstruction demonstrates solid union of the osteosynthesis site.

Complications and Clinical Results

Allografts used for tumor reconstruction have a high rate of early complications. Cumulative complication rates approach 50% in some series, with most patients requiring additional surgery.19 Infection, fracture, joint instability, and nonunion have vexed allograft reconstruction for 30 years.19-26 Immunologic complications27 and risk of disease transmission28 are of lesser concern but do exist. With osteoarticular allografts, osteoarthritis becomes manifest at 5 to 10 years in 15% of patients and is best treated with a resurfacing arthroplasty.19 Tumor recurrence, infection, and fracture are the most devastating complications and account for more than 85% of allograft failures.19 Most of the clinical outcome studies pertaining to the use of massive allografts come from the Orthopaedic Oncology Unit at Massachusetts General Hospital, with a series of more than 1,100 allograft reconstructions.19-24

Lord et al20 reported on the incidence, nature, and treatment of infections in bone allografts. A retrospective review of 283 patients with more than 2 years of follow-up revealed an infection rate of 11.7% (33/283). Gram-positive organisms, particularly Staphylococcus epidermidis, were the most common pathogens. Risk factors reflected those of a population treated by wide resection of soft tissue and bone, chemotherapy, and radiation therapy. Wound complications are the most common problem and were associated with early infection; additional surgical intervention is the most common risk factor for late infection. Eighty-two percent of infected cases (27/33) were considered failures and required amputation or removal of the allograft to control infection. Salvage is sometimes possible, but it requires an aggressive approach involving resection of the infected allograft, implantation of a spacer or external fixation, intravenous antibiotics, oral antibiotics for extended periods, and reimplantation of a new allograft or conversion to a metallic endoprosthesis.

In a retrospective review of 274 allograft recipients after limb-sparing tumor resections with a minimum follow-up of 4 years, Berrey et al21 reported a fracture incidence of 16% (43/274). There were no major distinguishing or predictive features in the allograft group between patients who had a fracture of the allograft and those who did not. There was a trend toward a higher incidence of nonunion in patients with fractures, but it was not statistically significant. The mean time to fracture was 28.6 months after the index surgery, with more than 70% of fractures occurring within 3 years. The results in this population after treatment approached those of patients who had never had a fracture. The mean time to union was 7.4 months (range, 4 to 14 months), with all but four fractures treated with an operation and most involving autogenous bone grafting. Weight bearing was restricted until radiographic union. The 43 patients underwent a total of 59 operations. The authors concluded that 9.3% of allograft shaft fractures (4/43) may heal with immobilization, but many require treatment with internal fixation and bone grafting. Several attempts may be necessary, and sometimes exchange of the allograft or conversion to a metallic endoprosthesis is necessary.

Berrey et al21 classified the fractures into three patterns. Type I fractures (2/43) were seen soon after surgery, with almost complete dissolution of the graft. These were thought to be secondary to an immune reaction to the allograft. Type II fractures (22/43) were through the shaft of the allograft, with a mean time to fracture of 27.6 months. Type III fractures (19/43) occurred at the articular surface of osteoarticular allografts at a mean of 31.6 months from surgery. These are best treated with a standard resurfacing total knee arthroplasty, when feasible.

In a larger series of allograft reconstructions,22 fractures occurred at a rate of 17.7% (185/1,046). Mean time to fracture was 3.2 years. There were 8 Berrey type I fractures, 114 type II, and 63 type III. Sixtyone fractures (33%) involved a screw hole at the end of the plate. Neither adjuvant chemotherapy nor radiation influenced the rate of allograft fracture.

Nonunion is another common complication in the postoperative course of allograft reconstructions. All allografts have at least one osteosynthesis site that, until healed, limits the amount of weight bearing permitted through the reconstructed limb. The location of the osteosynthesis affects the healing potential. Diaphyseal-to-diaphyseal osteosynthesis sites have a higher risk of delayed union or nonunion than do metaphyseal-to-metaphyseal sites. Supplemental autograft and stronger internal fixation are recommended at osteosynthesis sites in an effort to decrease the rate of nonunion. In the future, bone morphogenetic proteins may play a similar role when allograft reconstruction is performed. Hornicek et al23 evaluated factors affecting nonunion of the allograft-host junction. Of 945 patients, 163 (17.3%) had a nonunion. Those receiving chemotherapy had twice the rate of nonunion. Two hundred sixty-nine additional surgeries were performed on these 163 patients. In 114 patients, treatment led to successful union. The percentage of failure increased as the number of surgical procedures increased. Despite treatment, 49 patients failed to demonstrate union of the osteosynthesis site.

In comparing different types of allografts, intercalary allografts have better clinical outcomes than do osteoarticular allografts, allograftprosthetic composites, and allografts used for arthrodesis6,19,24 (Fig. 3). Ortiz-Cruz et al24 reviewed 104 intercalary allografts done over an 18-year period (median follow-up, 5.6 years). Eighty-four percent (87/104) were considered successful, with retention of allograft and normal extremity function. Infection, fracture, stage of disease, and adjuvant therapy all had adverse effects on graft survival. Fifteen reconstructions failed, most within 3 to 4 years. Four were salvaged with a second allograft, three by another reconstruction technique; eight required amputation (two for local recurrence). With two osteosynthesis sites, nonunions might be expected to be a common concern with intercalary allografts. Thirty-one of the 104 allografts (30%) failed to unite at one or both junctions within 1 year, but only seven remained ununited (and were considered failures) after additional surgical intervention. Eighty-one additional surgical procedures were needed to achieve satisfactory function in 92% of these patients. Different modes of internal fixation were used; plate fixation spanning both osteosynthesis sites was found to be superior to the use of two shorter plates at either end of the allograft.

Figure 3
Figure 3:
A 21-year-old man presented with thigh pain. Anteroposterior (A) and lateral (B) radiographs of the femur show a permeative diaphyseal lesion with cortical erosion (white arrows). C, Coronal T2-weighted MRI scan demonstrates the diaphyseal lesion (black arrow). Open biopsy confirmed Ewing's sarcoma. D, Postoperative anteroposterior radiograph of the femur after wide tumor resection and intercalary allograft reconstruction stabilized with a statically locked intramedullary nail.

Mankin et al19 found similar results in their review of 718 allograft transplantations (mean follow-up, 78 months). Intercalary allografts yielded the greatest satisfaction, with excellent or good outcomes in 84% of cases. Osteoarticular allografts, allograft-prosthetic composites, and allograft arthrodeses had excellent and good outcome rates of 73%, 77%, and 54%, respectively, although they often required additional surgical procedures to achieve these outcomes.

Regional Considerations

Most allograft reconstructions are for the femur, and the results of large studies primarily reflect the outcome for this region. Hornicek et al25 reviewed the largest series to date of proximal tibial osteoarticular allografts, consisting of 38 reconstructions (38 patients) in 15 years. Fifty-five percent of the patients experienced one or more complications, which were managed with multiple subsequent procedures. Three amputations were done for deep infections. About one third of the patients required removal of the original allograft and reconstruction with a new allograft or conversion to metallic prosthesis. Ultimately, 66% had a good or excellent functional result. The study also compared outcomes and complications between the subset of patients treated with chemotherapy, radiation therapy, or both to those without adjuvant therapy. The only significant (P < 0.05) difference observed between the groups was the higher incidence of fracture in the patients treated with chemotherapy.

Proximal humeral osteoarticular allografts are an attractive option because of their potential for soft-tissue reconstruction, healing, and function. O'Connor et al18 reported on eight patients treated with this technique, a subset of 57 patients who underwent limb-sparing tumor resections and various forms of reconstruction. There was no nonunion or cases of instability, but half of the patients experienced subchondral fractures and collapse of the articular surface. Three of these four patients were asymptomatic; the other was treated with conversion to a prosthesis. Functional rating averaged 71% by the MSTS system, with patients least satisfied with function and positioning of the hand. Compared with endoprosthetic reconstruction, osteoarticular allografts resulted in superior function after intra-articular resection of the proximal humerus. Shoulder arthrodesis using an intercalary allograft combined with plate fixation and vascularized fibular grafting is an excellent method of reconstruction after extra-articular resection of the proximal humerus.

Getty and Peabody26 reported similar results in 16 patients who underwent osteoarticular allograft reconstruction after intra-articular resection of the proximal humerus. At a mean follow-up of 47 months, the mean MSTS functional evaluation score was 70%. Deterioration was noted to continue with time from surgery. The authors have stopped doing the procedure because of the unacceptable rates of epiphyseal fragmentation (4/16), instability (11/16), fracture (4/16), and infection (1/16).

Allograft-Prosthetic Composites

Clearly there are benefits and inherent drawbacks to either allograft or endoprosthetic reconstruction. By combining the two methods or using an allograft-prosthetic composite, the surgeon can tailor the procedure to help diminish the inherent risks encountered when either reconstruction is used alone. The composite helps restore as much bone stock as possible and offers joint stability that is often difficult to obtain with osteoarticular allograft reconstruction. By resurfacing the allograft bone with an implant, cartilage degradation is no longer a potential problem 5 to 10 years after reconstruction (Fig. 4).

Figure 4
Figure 4:
A 14-year-old boy presented with knee pain of 3 months' duration and an enlarging mass. A, Anteroposterior radiograph of the knee demonstrates an eccentric lytic proximal tibial metaphyseal lesion with cortical destruction and soft-tissue extension (arrow). Biopsy confirmed osteosarcoma. Anteroposterior (B) and lateral (C) radiographs demonstrate skeletal reconstruction with a rotating-hinge knee prosthesis and proximal tibial allograft-prosthetic composite stabilized with a dynamic compression plate.

Gitelis and Piasecki14 performed 11 hip and 10 knee reconstructions and 1 elbow reconstruction in 22 patients (mean follow-up, 45 months). Mean MSTS functional score was 94.3%. Five patients had a nonunion, four of which healed after bone grafting; one was converted to a megaprosthesis. There were no dislocations. Graft resorption did not occur in this small number of patients, and no revisions were done for implant loosening.

Techniques With Special Indications

Vascularized Bone Grafts

Vascularized bone grafts can be taken from the iliac crest, rib, scapula, or fibula. Of these options, only vascularized fibular grafts are suited for the large skeletal defects left after wide resection of a malignant bone tumor. Compared with allografts, vascularized autografts offer a more rapid incorporation, stronger initial construct secondary to graft hypertrophy, and absence of immunologic problems. Vascular grafts change not the pattern of bony repair but rather the rate of repair. Final maturation and hypertrophy of grafts is consistent with Wolff's law.29 External fixation is preferable to plate fixation because it maximizes these stresses and allows for greater hypertrophy and ultimate strength of the graft. Ideally suited for children and young adults, this method of biologic reconstruction has the potential to be enduring without need for revision surgery later in life. Vascularized autografts also are used with the poorly vascularized tumor bed commonly found in previously irradiated tissue and when a delay in osteosynthesis healing is anticipated secondary to adjuvant therapy with radiation, chemotherapy, or both. The main disadvantages of vascularized autograft are the increased surgery time, surgical site morbidity, and size limitations.

Few published reports focus on vascularized autograft in tumor reconstruction, and long-term outcome data are lacking.30 Hsu et al31 reviewed a consecutive series of 30 patients who underwent skeletal reconstruction by vascularized fibular transfer after resection of primary bone tumors. Mean follow-up was 36 months (range, 24 to 85 months), with union achieved in 90% (27/30) at an average of 7.6 months. The mean fibular graft length was 18.9 cm (range, 10 to 30 cm). Functional results were evaluated in 24 patients, with 9 excellent, 7 good, 6 fair, and 2 poor results. When used for intercalary grafts (14 patients), the functional results were better than those seen with arthrodesis procedures (10 patients). There was a high complication rate (50%), but many were managed nonsurgically and resolved without greatly affecting the final outcome. Complications included three nonunions, three deep infections, three stress fractures, two local recurrences, and an assortment of soft-tissue complications.


Arthrodesis creates a stable, painless, durable limb. Indications for arthrodesis are extra-articular joint resection or extensive muscle resection with lack of remaining muscle to power the joint, or when the desire for joint stability is paramount. The two most common regions for this technique are the knee and shoulder.

Knee arthrodesis can be accomplished with allografts, nonvascularized autografts, vascularized rotational fibular grafts, external fixation with bone transport, or some combination of these techniques. Fixation is achieved with either compression plating or intramedullary nailing. Intramedullary fixation is favored for arthrodesis because of a decreased rate of graft fracture and nonunions. The knee is aligned in 10° to 15° of flexion and 0° to 5° of valgus. In the skeletally mature individual, the limb is shortened 1 to 2 cm to allow for foot clearance during the gait cycle. In skeletally immature patients, the limb may be lengthened with the grafting technique; the expectation is that the contralateral normal limb will continue to grow, with the result that limb length at skeletal maturity on the operated side will be equal or slightly shorter. The procedure is associated with a high rate of complications (approximately 50%), including all of the inherent risks of allograft reconstruction. Despite this complication rate, however, most patients achieve successful union and have a durable, functional limb.32-34

The shoulder joint is challenging to reconstruct, given the extreme range of motion and lack of inherent static stability. The few published reports that discuss shoulder arthrodesis in tumor reconstruction have small numbers of patients (5 to 10). Notable rates of infection, fracture, and nonunion exist, as with knee arthrodesis.32 Most patients obtain stable fusion, allowing satisfactory function of the upper extremity.18,32

Wolf et al33 reviewed the longterm results in 73 patients who, from 1967 to 1985, underwent resection arthrodesis of the knee with autogenous grafts. Forty patients followed for more than 10 years formed the basis of the evaluation. Intramedullary rods were used to stabilize hemicortical femoral or tibial allografts and nonvascularized autogenous fibular grafts to the native femur and tibia. A high incidence of complications (52%) was evident, yet most patients eventually achieved a successful outcome with preservation of the limb, and 86% were independent ambulators at long-term follow-up. The most common complications were graft fatigue fracture (51%), delayed union (23%), rod migration or rod fracture (25%), peroneal nerve palsy (8%), and infection and wound problems (23%). Thirty-seven of the 40 patients (93%) achieved a solid reconstruction. Two patients required an above-knee amputation, one for infection and the other for local tumor recurrence. One additional patient sustained a comminuted traumatic fracture of the fusion and was treated with an allograft arthrodesis. The reconstructions proved to be durable, and patient satisfaction and function remained high for decades, with an average MSTS functional score of 77%.

Weiner et al34 evaluated 39 patients treated with resection arthrodesis done with an intercalary allograft fixed with an intramedullary nail. In 31 patients, this procedure was the index reconstruction; in eight, it was done after failure of a different type of tumor reconstruction. Proximal and distal osteosynthesis sites both healed, and function was satisfactory in 32 patients. Nonunion occurred in seven patients (one junction in six patients, both junctions in one). The nonunion was healed in five of the seven treated with bone grafting, repeat internal fixation, or exchange allografting; two patients went on to above-knee amputation. An additional patient from the study underwent above-knee amputation for local recurrence. Fatigue fracture of the allograft occurred in five patients, all within metaphyseal bone. Overall, the rate of complication was lower than that experienced with autograft arthrodesis.

Expandable Prostheses

Expandable prostheses were developed in an attempt to overcome anticipated limb-length discrepancies in the growing child treated with limb-sparing surgery. The first approach includes the Lewis Expandable Adjustable Prosthesis (LEAP).35 Expansion of the LEAP is achieved by a modified Jacob's chuck mechanism. Rotation of the outer sleeve of the prosthesis engages the threads on the inner shaft, thereby increasing the length of the prosthetic shaft with each revolution. An average lengthening is 1.5 to 2 cm; the average overall extension capability of a LEAP is 6 to 9 cm. Collapse of the expansion mechanism, observed early in its use, has been addressed with the use of spacer rings. Problems with titanium debris and fatigue failure of expandable prostheses led to the alternative use of modular systems. These systems use a Morse taper locking system to connect segmental parts. Exchange of intercalary segments can be performed as the patient grows, adding 2 cm to the length of the segment replaced.

When extremity length discrepancy reaches approximately 2 cm or more, an expansion procedure is indicated. The original incision is used, and the pseudocapsule around the prosthesis is excised to prevent problems with joint stiffness after lengthening and to relieve tension on the neurovascular bundle. Eckardt et al36 reported on their 14-year experience with 32 expandable prostheses. Nineteen of the 32 patients (59%) survived, with a median follow-up of 105 months. Sixteen of the patients (50%) did not undergo an expansion because of death, amputation, or short duration of followup. The remaining 16 patients underwent 32 expansion procedures, to a maximum of 9 cm, without infection. The average time from implantation to the first lengthening was 19 months. Most of the lengthenings were 1.5 to 2 cm. More than 50% of the patients had at least one complication, most frequently aseptic loosening or failure of the prosthesis, collapse of the LEAP, temporary nerve palsy, or flexion contractures. The average MSTS ratings were good to excellent at the knee, fair to good at the hip, and fair about the shoulder. The authors concluded that, in children and the skeletally immature, rehabilitation can be problematic; early loss of joint motion and fixed flexion contractures can occur. With advances in technology, a noninvasive prosthetic lenghtening mechanism is now being evaluated clinically.


Rotationplasty, another reconstruction option in the skeletally immature patient, can be done after wide resection about the knee when the sciatic nerve can be preserved. The tibia is rotated 180° and fused to the femur, with the ankle joint placed at the level of the contralateral knee. The procedure creates a functional below-knee amputation; outcome far exceeds that associated with above-knee ablation. The main indication for this technique is in a very young child with an extensive malignant bone tumor and several years of growth remaining. Rotationplasty also may be used in adults when soft-tissue coverage is inadequate after extensive tumor resection. In general, the procedure is associated with a low incidence of complications and a highly functional and durable extremity reconstruction.37 Advantages include the maintenance of growth and a functioning “knee” joint, the ability to tailor the procedure to obtain limblength equality at skeletal maturity without further operations, an energyefficient gait pattern, and avoidance of the problems of neuromas and stump breakdown seen with amputations. With the advent of expandable prostheses in the late 1980s, however, the indications for rotationplasty have narrowed. A full presurgical discussion is essential so that the patient and family understand the advantages and cosmetic appearance of the reconstructed limb, thus limiting postoperative dissatisfaction or psychological problems. Meetings with other patients who have had the procedure are beneficial.

Kotz37 reviewed the results of 40 patients treated with rotationplasty between 1976 and 1988. Thirty were followed for more than 3 years. No patient developed local recurrence; six died from metastatic disease, and the remaining 24 were tumor free. All patients were prosthetic ambulators without additional supports, and most participated in sports. Functional evaluation revealed 68% excellent, 28.5% good, 3.5% fair, and no poor results according to the system of Enneking et al.38 Complications of rotationplasty include postoperative vascular occlusion, pseudarthrosis between femur and tibia, nerve palsies, rotational malalignment, and diffuse osteopenia in the distal limb bones. The high level of function achieved by most patients far outweighs the appearance of the limb, limiting the psychological problems associated with rotationplasty.37,39 Long-term follow-up studies have shown the high durability of rotationplasty, with continued excellent or good results at 8 years.39

Limb Lengthening

The Ilizarov bone transport procedure and other techniques of limb lengthening can be used to regain bone length after resection and reconstruction. Limb lengthening by distraction osteogenesis or bony transport has limited utility after resection of malignant bone tumors when used as the primary reconstruction technique. The large osseous defect is difficult to replace and requires extended periods of treatment, which are associated with significant complications. Frequently, the final functional result with this technique is poor.40 Limblengthening procedures are better suited as adjuncts to other methods of reconstruction or for smaller defects.

Limb-Sparing Procedures Compared With Amputation

Survival and Local Recurrence Rates

In comparing limb-sparing procedures with amputation, the points selected as outcome measures are important. The primary goal of any oncologic procedure is local tumor control to diminish local recurrence and improve overall survival. Limb-preserving procedures have not decreased overall survival rates4,5,11,19,41,42 (Table 3). Local recurrence associated with limb-sparing resection and reconstruction is slightly greater than it is after amputation or joint disarticulation, but this has not been found ultimately to affect patient survival.4,41 In the largest series to date comparing limb-sparing surgery with amputation in 227 patients with osteosarcoma of the distal femur, Rougraff et al41 found local recurrence in 8 of 73 patients treated with limb preservation, in 9 of 115 patients treated with above-knee amputation, and in none of 39 patients treated with hip disarticulation.

Table 3
Table 3:
Limb Sparing Versus Amputation for Extremity Bone Sarcomas

Functional Outcome

The MSTS system for assessing the function of reconstructive procedures (Table 2) is designed to allow for comparison of results.16 Functional outcome studies comparing limb preservation and amputation have inherent limitations, however, including the inability to randomize treatment and the subjective nature of important outcome measures. Most functional outcome measures favor nonarthrodesis procedures because range of motion is measured. Most of the studies designed to assess functional outcome focus on patients with perigeniculate tumor resections. Good and bad outcomes occur with whatever procedure is used, and there is little difference in quality-of-life outcomes between limb-sparing and amputative surgeries.43

Rougraff et al41 found that their limb-preservation group had higher functional scores than did the group treated with amputation but that the limb-preservation patients frequently required additional surgical procedures to reach peak function. Renard et al42 found functional results to be significantly (P = 0.0001) better after limb-saving surgery compared with ablative therapy; however, complications were three times more common in the limbsaving cohort.

Otis et al44 studied the energy cost during gait by measuring oxygen consumption in 14 patients with custom-made knee prostheses and 12 patients who had had aboveknee amputations and been fitted with an artificial limb. They concluded that prosthetic reconstruction provides superior function because these patients had a lower energy cost during gait. In contrast, the patients studied by Harris et al45 functioned similarly and walked with comparable velocity, efficiency, and rate of oxygen consumption whether they had had an amputation, arthrodesis, or arthroplasty. The patients treated with endoprostheses lived more sedentary lives and were the most protective of the limb.

Kawai et al46 evaluated clinical outcomes, length of resection, and energy cost of walking after prosthetic knee replacement for malignant tumors of the distal femur. The mean free-walking velocity was 79% of normal, reflecting a decrease in both cadence and stride length. Also, mean energy cost during walking was 35% greater than that of normal control subjects and correlated with the percentage of femur that had been resected. Hillmann et al47 reported on 67 patients with malignant tumors of the distal femur or proximal tibia who were treated by rotationplasty or endoprosthetic reconstruction. Patients with a rotationplasty had a mean MSTS functional score of 80% (24/30) compared with 83.3% (25/30) for patients treated with an endoprosthesis. Those with a rotationplasty had fewer restrictions in their daily activities and required ambulatory assistive devices less frequently than did those reconstructed with a prosthesis. Finally, McClenaghan et al48 compared oxygen consumption in patients treated with above-knee amputation, arthrodesis, or rotationplasty. Patients treated by rotationplasty walked the most efficiently.

Psychological Factors

Whether limb-sparing surgery offers a psychological outcome advantage compared with amputation for extremity sarcomas has yet to be demonstrated because no long-term prospective or comparative studies have been done.5,49 In a small, retrospective, one-time psychological assessment of patients treated for lower extremity sarcomas, no differences between amputation and limb sparing were found regarding cognitive capacity, mood, body image, global physical functioning, global adjustment to illness and surgery, and lifetime prevalence of psychological disorders before or after surgery.49 Most patients adjust well to both the disease and the required surgical treatment if they have no premorbid psychological disorders.5


Grimer et al43 demonstrated that endoprosthetic reconstruction is less expensive than amputation, based on a 1997 cost analysis and 20-year follow-up. Their formula takes into account the projected need for revision surgery, based on rates of aseptic loosening of 2.5% per year and on rates of other causes of early failure (eg, infection, implant failure) of 1.5% per year.

Future Directions

As experience is gained, the ability to amend methods of treatment to improve outcomes will increase. Endoprostheses continue to be improved. Long-term results of modular rotating-hinge knee components are anticipated to yield better results than those of simplehinge knee prostheses. A few institutions have had experience with an endoprosthesis that is lengthened noninvasively by the application of external electromagnetic force, and development continues. Soft-tissue reattachment and ligament reconstruction are easier to perform because of the use of osteoarticular allografts. With the development of tendon-attachment devices or enhanced tendon-anchorage devices for prostheses, the indications and outcomes for proximal tibial prostheses are likely to broaden and improve.3 Better stability around the hip and shoulder after prosthetic reconstructions fit with these special devices should improve functional results in these regions. Extracortical bone-bridging fixation will likely improve the longevity of prostheses by walling off the prosthesis-bone interface and adding additional points of fixation to improve construct strength.3,9

For reconstructions using allografts, bone morphogenetic proteins may decrease nonunion rates. A better understanding of allograft biology, as well as either closer matching of allografts to recipients or modulation of immune responses, may decrease the presumed consequences of allograft reconstruction. Using more vascularized bone grafts and combining them with allografts may help reduce or address rates of nonunion and fracture. The continued vigilant use of adequate soft-tissue coverage in reconstruction procedures, which reduced early postoperative wound complications, likely will decrease the incidence of late deep wounds.


The surgical management of malignant bone tumors of the extremities presents many challenges. With advances in chemotherapy, radiographic imaging, and reconstructive surgery, most patients with these rare tumors now can be offered limb-sparing surgery. Osteoarticular allografts, modular prostheses, or composites of these two approaches form the basis for most current reconstruction efforts. However, amputation still plays an important role and offers a standard to which other approaches must be compared. Functional outcome and patient satisfaction appear to be at least as good, and probably better, after skeletal reconstruction than after amputation. However, the surgical treatment regimen associated with limb-sparing procedures is also associated with significant complications and requires extensive rehabilitation. Outcomes should continue to improve as advances are made in surgical technique, implant design, autogenous bone allograft biology, and postoperative management.


1. Eilber FR, Eckhardt J, Morton DL: Advances in the treatment of sarcomas of the extremity: Current status of limb salvage. Cancer 1984;54(11 suppl): 2695-2701.
2. Enneking WF: An abbreviated history of orthopaedic oncology in North America. Clin Orthop 2000;374:115-124.
3. Choong PF, Sim FH: Limb-sparing surgery for bone tumors: New developments. Semin Surg Oncol 1997;13:64-69.
4. Sluga M, Windhager R, Lang S, Heinzl H, Bielack S, Kotz R: Local and systemic control after ablative and limb sparing surgery in patients with osteosarcoma. Clin Orthop 1999;358:120-127.
5. Simon MA: Limb salvage for osteosarcoma in the 1980s. Clin Orthop 1991; 270:264-270.
6. Hornicek FJ, Gebhardt MC, Sorger JI, Mankin HJ: Tumor reconstruction. Orthop Clin North Am 1999;30:673-684.
7. Horowitz SM, Glasser DB, Lane JM, Healey JH: Prosthetic and extremity survivorship after limb salvage for sarcoma: How long do the reconstructions last? Clin Orthop 1993;293:280-286.
8. Malawer MM, Chou LB: Prosthetic survival and clinical results with use of large-segment replacements in the treatment of high-grade bone sarcomas. J Bone Joint Surg Am 1995;77: 1154-1165.
9. Damron TA: Endoprosthetic replacement following limb-sparing resection for bone sarcoma. Semin Surg Oncol 1997;13:3-10.
10. Unwin PS, Cannon SR, Grimer RJ, Kemp HB, Sneath RS, Walker PS: Aseptic loosening in cemented custommade prosthetic replacements for bone tumours of the lower limb. J Bone Joint Surg Br 1996;78:5-13.
11. Eckardt JJ, Yang RS, Ward WG, Kelly C, Eilber FR: Endoprosthetic reconstruction for malignant bone tumors and nonmalignant tumorous conditions of bone, in Stauffer RN, Erlich MG, Fu FH, Kostuik JP, Manske PR, Sim FH (eds): Advances in Operative Orthopaedics. St. Louis, MO: Mosby, 1995, vol 3, pp 61-83.
12. Zehr RJ, Enneking WF, Scarborough MT: Allograft-prosthesis composite versus megaprosthesis in proximal femoral reconstruction. Clin Orthop 1996;322:207-223.
13. Kabukcuoglu Y, Grimer RJ, Tillman RM, Carter SR: Endoprosthetic replacement for primary malignant tumors of the proximal femur. Clin Orthop 1999;358:8-14.
14. Gitelis S, Piasecki P: Allograft prosthetic composite arthroplasty for osteosarcoma and other aggressive bone tumors. Clin Orthop 1991;270:197-201.
15. Kawai A, Muschler GF, Lane JM, Otis JC, Healey JH: Prosthetic knee replacement after resection of a malignant tumor of the distal part of the femur: Medium to long-term results. J Bone Joint Surg Am 1998;80:636-647.
16. Enneking WF, Dunham W, Gebhardt MC, Malawar M, Pritchard DJ: A system for the functional evaluation of reconstructive procedures after surgical treatment of tumors of the musculoskeletal system. Clin Orthop 1993; 286:241-246.
17. Grimer RJ, Carter SR, Tillman RM, et al: Endoprosthetic replacement of the proximal tibia. J Bone Joint Surg Br 1999;81:488-494.
18. O'Connor MI, Sim FH, Chao EY: Limb salvage for neoplasms of the shoulder girdle: Intermediate reconstructive and functional results. J Bone Joint Surg Am 1996;78:1872-1888.
19. Mankin HJ, Gebhardt MC, Jennings LC, Springfield DS, Tomford WW: Long-term results of allograft replacement in the management of bone tumors. Clin Orthop 1996;324:86-97.
20. Lord CF, Gebhardt MC, Tomford WW, Mankin HJ: Infection in bone allografts: Incidence, nature, and treatment. J Bone Joint Surg Am 1988;70:369-376.
21. Berrey BH Jr, Lord CF, Gebhardt MC, Mankin HJ: Fractures of allografts: Frequency, treatment, and end-results. J Bone Joint Surg Am 1990;72:825-833.
22. Sorger JI, Hornicek FJ, Zavatta M, et al: Allograft fractures revisited. Clin Orthop 2001;382:66-74.
23. Hornicek FJ, Gebhardt MC, Tomford WW, et al: Factors affecting nonunion of allograft-host junction. Clin Orthop 2001;382:87-98.
24. Ortiz-Cruz E, Gebhardt MC, Jennings LC, Springfield DS, Mankin HJ: The results of transplantation of intercalary allografts after resection of tumors: A long-term follow-up study. J Bone Joint Surg Am 1997;79:97-106.
25. Hornicek FJ Jr, Mnaymneh W, Lackman RD, Exner GU, Malinin TI: Limb salvage with osteoarticular allografts after resection of proximal tibia bone tumors. Clin Orthop 1998;352:179-186.
26. Getty PJ, Peabody TD: Complications and functional outcomes of reconstruction with an osteoarticular allograft after intra-articular resection of the proximal aspect of the humerus. J Bone Joint Surg Am 1999;81:1138-1146.
27. Friedlaender GE: Bone allografts: The biological consequences of immunological events. J Bone Joint Surg Am 1991; 73:1119-1122.
28. Friedlaender GE: Appropriate screening for prevention of infection transmission by musculoskeletal allografts. Instr Course Lect 2000;49:615-619.
29. Shaffer JW, Field GA, Goldberg VM, Davy DT: Fate of vascularized and nonvascularized autografts. Clin Orthop 1985;197:32.
30. Brown KL: Limb reconstruction with vascularized fibular grafts after bone tumor resection. Clin Orthop 1991;262: 64-73.
31. Hsu RW-W, Wood MB, Sim FH, Chao EYS: Free vascularised fibular grafting for reconstruction after tumour resection. J Bone Joint Surg Br 1997;79:36-42.
32. Scarborough MT, Helmstedter CS: Arthrodesis after resection of bone tumors. Semin Surg Oncol 1997;13:25-33.
33. Wolf RE, Scarborough MT, Enneking WF: Long-term followup of patients with autogenous resection arthrodesis of the knee. Clin Orthop 1999;358:36-40.
34. Weiner SD, Scarborough M, Vander Griend RA: Resection arthrodesis of the knee with an intercalary allograft. J Bone Joint Surg Am 1996;78:185-192.
35. Kenan S, Bloom N, Lewis MM: Limbsparing surgery in skeletally immature patients with osteosarcoma: The use of an expandable prosthesis. Clin Orthop 1991;270:223-230.
36. Eckardt JJ, Kabo JM, Kelley CM, et al: Expandable endoprosthesis reconstruction in skeletally immature patients with tumors. Clin Orthop 2000;373:51-61.
37. Kotz R: Rotationplasty. Semin Surg Oncol 1997;13:34-40.
38. Enneking WF, Spanier SS, Goodman MA: A system for the surgical staging of musculoskeletal sarcoma. Clin Orthop 1980;153:106-120.
39. Hanlon M, Krajbich JI: Rotationplasty in skeletally immature patients: Longterm followup results. Clin Orthop 1999;358:75-82.
40. Ozaki T, Nakatsuka Y, Kunisada T, et al: High complication rate of reconstruction using Ilizarov bone transport method in patients with bone sarcomas. Arch Orthop Trauma Surg 1998;118: 136-139.
41. Rougraff BT, Simon MA, Kneisl JS, Greenberg DB, Mankin HJ: Limb salvage compared with amputation for osteosarcoma of the distal end of the femur: A long-term oncological, functional, and quality-of-life study. J Bone Joint Surg Am 1994;76:649-656.
42. Renard AJ, Veth RP, Schreuder HW, van Loon CJ, Koops HS, van Horn JR: Function and complications after ablative and limb-salvage therapy in lower extremity sarcoma of bone. J Surg Oncol 2000;73:198-205.
43. Grimer RJ, Carter SR, Pynsent PB: The cost-effectiveness of limb salvage for bone tumours. J Bone Joint Surg Br 1997;79:558-561.
44. Otis JC, Lane JM, Kroll MA: Energy cost during gait in osteosarcoma patients after resection and knee replacement and after above-the-knee amputation. J Bone Joint Surg Am 1985;67:606-611.
45. Harris IE, Leff AR, Gitelis S, Simon MA: Function after amputation, arthrodesis, or arthroplasty for tumors about the knee. J Bone Joint Surg Am 1990;72:1477-1485.
46. Kawai A, Backus SI, Otis JC, Healey JH: Interrelationships of clinical outcome, length of resection, and energy cost of walking after prosthetic knee replacement following resection of a malignant tumor of the distal aspect of the femur. J Bone Joint Surg Am 1998;80:822-831.
47. Hillmann A, Hoffmann C, Gosheger G, Krakau H, Winkelmann W: Malignant tumor of the distal part of the femur or the proximal part of the tibia: Endoprosthetic replacement or rotationplasty. Functional outcome and quality-of-life measurements. J Bone Joint Surg Am 1999;81:462-468.
48. McClenaghan BA, Krajbich JI, Pirone AM, Koheil R, Longmuir P: Comparative assessment of gait after limb-salvage procedures. J Bone Joint Surg Am 1989;71:1178-1182.
49. Weddington WW Jr, Segraves KB, Simon MA: Psychological outcome of extremity sarcoma survivors undergoing amputation or limb salvage. J Clin Oncol 1985;3:1393-1399.
© 2003 by American Academy of Orthopaedic Surgeons